distribution and ecology of the broad-toothed rat in …€¦ · determine the distribution,...

DISTRIBUTION AND ECOLOGY OF

THE BROAD-TOOTHED RAT

IN THE ACT

RICHARD MILNER, DANSWELL STARRS,

GREG HAYES AND MURRAY EVANS

Conservation Research

Technical Report 35

October 2016

Technical Report 35

Distribution and ecology of the

Broad-toothed Rat in the ACT

Richard Milner*, Danswell Starrs^, Greg Hayes* and Murray Evans*

*Conservation Research, Environment and Planning Directorate,

ACT Government, Canberra, ACT 2601

^Evolution, Ecology and Genetics, Research School of Biology,

Australian National University, Canberra, ACT 0200

Conservation Research

Environment Division

Environment and Planning Directorate

October 2016

ISBN: 978-1-921117-50-0 Technical Report Series: ISSN 1320-1069: 35 © Australian Capital Territory, Canberra, 2016 Information contained in this publication may be copied or reproduced for study, research, information or educational purposes, subject to appropriate referencing of the source. This document should be cited as: Milner R, Starrs D, Hayes G and Evans M. 2016. Distribution and ecology of the broad-toothed rat in the ACT. Technical Report 35. Environment and Planning Directorate, ACT Government, Canberra. http://www.environment.act.gov.au Telephone: Access Canberra 13 22 81 Disclaimer The views and opinions expressed in this report are those of the authors and do not necessarily represent the views, opinions or policy of funding bodies or participating member agencies or organisations. Front cover: Broad-toothed Rat. G A Hoye © Australian Museum.

Distribution and ecology of the Broad-toothed Rat i

Contents

Summary ..................................................................................................................................................ii

1 Introduction ..................................................................................................................................... 1

2 Background ...................................................................................................................................... 2

2.1 Description and Distribution ................................................................................................... 2

2.2 Status ....................................................................................................................................... 2

2.3 Habitat ..................................................................................................................................... 3

2.4 Diet .......................................................................................................................................... 3

2.5 Population dynamics ............................................................................................................... 3

2.6 Breeding biology ...................................................................................................................... 4

2.7 Threatening processes ............................................................................................................ 4

3 Methods .......................................................................................................................................... 5

3.1 Study sites ............................................................................................................................... 5

3.2 Survey method ........................................................................................................................ 5

3.3 Statistical Analyses .................................................................................................................. 6

4 Results ............................................................................................................................................. 8

4.1 Bogs ......................................................................................................................................... 8

4.2 Distribution of M. fuscus ......................................................................................................... 8

4.3 Habitat preference of M. fuscus .............................................................................................. 8

4.4 Feral animal disturbance ......................................................................................................... 9

4.5 Observer bias ........................................................................................................................... 9

5 Discussion ...................................................................................................................................... 18

5.1 Distribution, status and abundance ...................................................................................... 18

5.2 ACT Habitat ............................................................................................................................ 18

5.3 Bog size .................................................................................................................................. 19

5.4 Vegetation composition ........................................................................................................ 19

5.5 Threats and Management recommendations ...................................................................... 19

5.6 Conclusion ............................................................................................................................. 22

Acknowledgements ............................................................................................................................... 22

References ............................................................................................................................................. 22

Appendix................................................................................................................................................ 25

ii Distribution and ecology of the Broad-toothed Rat

Summary

The Broad-toothed Rat Mastacomys fuscus is a small to medium sized rodent of the family Muridae.

Within the ACT the species is thought to occur in subalpine swamps and the associated riparian

heathland, sedgeland and grassland environments above 1400 m. M. fuscus is recognised nationally

as a declining species endemic to alpine and sub-alpine regions of south-eastern Australia, including

Tasmania, Victoria, NSW and the ACT. The species is identified as being at extreme risk to global

warming due to shifts in composition and distribution of alpine vegetation communities. Climate

change has been identified as a key threatening process to the long-term viability of the species. In

NSW M. fuscus is classified as vulnerable, with one of the two populations, Barrington Tops, classified

as endangered. In Victoria the species is classified as near threatened and in the ACT, where only

limited information is available on the species, M. fuscus has no special protection status.

The purpose of this study was to determine the presence or absence of the M. Fuscus in known

habitat for the species (sphagnum moss bogs and fens) in the ACT, and to identify key ecological

attributes that determine the presence of the species. A key aim of this study was to establish a long

term monitoring program to assess long-term trends in the species’ abundance and distribution,

which may be affected under projected climate change scenarios.

Fourteen ACT alpine bogs were surveyed across Namadgi National Park, which covers the southern

half of the ACT and adjoins Kosciusko National Park. Scat searches were undertaken in 27-132

quadrats per bog to determine the species presence/absence, relative abundance and habitat

preference. For each quadrat, information on M. fuscus presence/absence, vegetation type, feral

animal disturbance and distance to drainage lines was recorded. Bog size and distance to the nearest

occupied bog was also recorded.

Results indicate M. fuscus presence across 13 of the 14 sites: Cotter Gap, Cheyenne, Rocks Flat,

Snowy Flat, Ginini West, Big Creamy, Ginini East, Tom Gregory, Little Bimberi, Rotten Swamp, Murray

Flat, Leura and Top Flat. The species showed a positive habitat preference for vegetation

communities dominated by heath, sedge and Poa, and was also identified in a further nine dominant

vegetation types, including Eucalyptus pauciflora (snowgum) woodland. Vegetation composition

varied with bog size and M. fuscus detection rate increased with increasing bog size. Distance from

drainage lines was also important, with detection rates increasing with proximity to drainage lines.

Finally the presence of feral animals such as foxes, rabbits and pigs significantly influenced detection

rates, with M. fuscus detection rate decreasing as signs of feral animal disturbance increased.

This study provides land managers with a baseline assessment of M. fuscus condition in the ACT

against which future changes can be assessed and key-threatening processes can be monitored.

The results of this study have also been published in:

Milner R, Starrs D, Hayes G and Evans M. 2014. Distribution and habitat preference of the broad-

toothed rat (Mastacomys fuscus) in the Australian Capital Territory, Australia. Australian

Mammalogy. http://dx.doi.org/10.1071/AM14031

Distribution and ecology of the Broad-toothed Rat 1

1 Introduction

The Broad-toothed Rat, Mastacomys fuscus, is a cryptic native rodent that inhabits montane,

subalpine and alpine areas, usually in close association with water bodies. In the ACT M. fuscus is

known to occur in sphagnum moss bogs, although surveys for the species have not been conducted

for over two decades and only limited information exists on its distribution and abundance. The

species is known to have declined in other jurisdictions, and is listed as threatened in NSW and

Victoria. Climate change has been identified as a key threatening process to the long-term viability of

the M. fuscus due to shifts in composition and distribution of alpine vegetation communities.

The broad goal of this study was to gain a better understanding of the distribution, abundance,

ecology and conservation status of M. fuscus in the ACT to support appropriate management

decisions for the species and its habitat.

Specific aims of study were to:

Determine the distribution, abundance and status of the Broad-toothed Rat Mastacomys

fuscus in the ACT.

Identify key ecological attributes that determine or influence the likelihood of the species

occurrence.

Establish a baseline and method for long term monitoring to assess the effects of climate

change on the abundance and distribution of the species.

2 Distribution and ecology of the Broad-toothed Rat

2 Background

2.1 Description and Distribution

The Broad-toothed Rat, Mastacomys fuscus, is a small to medium sized rodent of the family Muridae

(Happold 2002). Adult head-body length ranges from 142–175 mm, tail length ranges from 100–130

mm and adult weight ranges from 95–145 g. The species is found across a range of altitudes from sea

level to 2200 m where habitat is suitable in the Dandenong and Otway Ranges, Victorian Alps, Snowy

Mountains, Barrington Tops and Tasmania;. Across Snowy Mountains and Barrington Tops

populations, the species habitat is characterised by mean annual temperatures 1000 mm (Green and Osborne 2003). While the species is currently restricted to

patches of apparently optimal habitat, fossil records indicate a more extensive distribution during the

Pleistocene when the climate was cooler (Ride 1956).

Within the ACT there is limited information on the distribution, abundance and ecology of M. fuscus.

Only three surveys in the ACT have successfully detected the species, all of these were undertaken

prior to 1990. In 1973 Eberhard and Schulz (1973) captured a single M. fuscus at Murrays Gap when

undertaking a vertebrate survey of the Cotter River catchment. In 1986, Lintermans (1986)

undertook a survey of M. fuscus across six ACT sites (Murrays Gap, Cheyenne Flats, Stockyard Saddle,

saddle below Mt Gingera, Mt Gingera and Snowy Flats) from which only one individual was captured

(at Murray Gap). From 1988 to 1989, Green and Osborne (2003) surveyed 25 areas within or

bordering the ACT with M. fuscus recorded within 12 of the sites: Hanging Flat, Murrays Gap, Mt

Murray, Ginini Flats, Cheyenne Flats, Stockyard Flats, Brumby Flats, 80 Acres, Blackfellows Gaps,

Rolling Ground Gap, Mt Bimberi and Mt Scabby. Runways and scats of M. fuscus have been observed

at Ginini Flats, Cheyenne Flats, Morass Flats and Snowy Flats (M Evans pers. obs.; D Whitfield pers.

obs.).

2.2 Status

M. fuscus is recognised nationally as a declining species endemic to alpine and sub-alpine regions of

south-eastern Australia, including Tasmania, Victoria, NSW and the ACT. In Victoria M. fuscus is listed

as Near Threatened, and in NSW the species is listed as Threatened (Vulnerable) though one of the

populations (Barrington Tops) is classified as Endangered. M. fuscus is not listed as threatened in the

ACT, where there is a paucity of information on the species.

M. fuscus is identified as being at extreme risk to global warming due to loss of habitat (Sphagnum

moss bogs and other wet seepage areas) resulting from shifts in the composition and distribution of

alpine vegetation communities (DECC NSW 2007), and accordingly climate change has been

identified as a key threatening process to the long-term viability of M. fuscus. Brereton and

colleagues (1995) predicted that with a 1oC rise in temperature the range of M. fuscus would

decrease by 36% and with a 3 oC rise in temperature it would decrease by 75%. In 2005, O’Brien

showed that extinction of M. fuscus at Barrington Tops is expected under almost any climatic

warming scenario.


2.3 Habitat

M. fuscus is associated with a variety of habitat types, including rocky outcrops, tall tussock

grasslands, reeds, sedges and low spreading heath shrubs, usually within 15 m of watercourses (K

Green pers. comm.). The species requires dense ground cover and an abundance of green grass,

including snow grasses (Poa spp). In the ACT, M. fuscus is apparently associated with subalpine

swamps and associated riparian heathland, sedgeland and grassland environments above 1000 m,

with a mean annual rainfall greater than 1000 mm and average temperature


winter (Bubela et al. 1991). Female home range size is determined by metabolic requirements such

as food quality and food density (Mace et al. 1983), while male home range size is also subject to

sexual selection for increased access to females, with home range decreasing as female reproductive

receptiveness decreases (Bubela et al. 1991). While M. fuscus do not appear to display territorial

defence, there is evidence that both sexes occupy an exclusive/solitary space that averages 802m2

within their home range during January to autumn (Bubela et al. 1991).

2.6 Breeding biology

M. fuscus display female-only parental care and are believed to be either polygamous or

promiscuous (Happold 2011). Mating occurs in mid-late October and is trigged by a gradual increase

in ambient temperature, snowmelt, decreased frequency of frost, increasing day length and the

growth of new grass (Bubela and Happold 1993). Breeding nests are built under dense shrubs,

mattered stems, boulders and fallen logs and are similar to overwinter nests – they have a single

entrance and a single internal cavity up to 30 cm in diameter (Happold 2011). Gestation period is 35

days (Calaby and Wimbush 1964) and all females synchronously give birth to their first litter in the

last week of November or first week of December (Happold 1989). Births of the second and

sometimes third litters are spread out over several weeks from December to early March and litter

size ranges from 1–4 (Bubela et al. 1991; 1993). Young M. fuscus start to eat grass at three weeks and

are weaned at five weeks (Happold 2011) and sexual maturity is reached at 8–10 months.

2.7 Threatening processes

Key threatening process recognised by the NSW Threatened Species Conservation ACT 1995 that are

relevant to M. fuscus include:

predation by European Red Fox Vulpes vulpes and the Feral Cat Felis catus;

competition and grazing by feral European Rabbits Oryctalagus cuniculus;

competition, disease transmission and habitat degradation by Feral Pigs (Sus scrofa);

invasion of exotic perennial grasses;

climate change resulting in loss of sub-alpine and alpine habitat, and;

spread of the plant root fungus Phytophthora cinnamomi.


3 Methods

3.1 Study sites

Potential habitat for M. fuscus was identified remotely using aerial photographs, vegetation maps

(i.e. sphagnum shrub bog, sod tussock Poa grassland, Poa-Danthonia grassland, and Restiad bog

ArcGIS layers – CPR GIS system), ACT Wildlife Atlas records, and expert knowledge (D Whitfield pers.

comm.). Fourteen study sites were identified, including: Cotter Source, Top Flat, Rotten Swamp, Big

Creamy, Rocks Flat, Cotter Gap, Tom Gregory, Murray Gap, Little Bimberi, Leura, Snowy Flat,

Cheyenne Flat, Ginini West and Ginini East.

3.2 Survey method

Surveys were undertaken in early autumn (March) 2013. Area-based searches for scats (faecal

pellets) were undertaken in 27–130 circular quadrats (2 m radius) per study site. Quadrats were

positioned 20 m apart along 5–12 straight-line transects. Line transects were positioned at right

angles to the streams and elevational gradient (i.e. transects cut across contours; Map 1) and were

placed 50 m apart for study sites/bogs > 35,000 m2 and 25 m apart for study sites/bogs


3.3 Statistical Analyses

The proportion of quadrats with M. fuscus scats was plotted against bog size (km2). Linear regression

was used to determine if a significant relationship existed between M. fuscus abundance and bog

size. Principle component analysis on the covariance matrix was used to explore variation in

vegetation types between bogs. Proportional cover of vegetation across 11 bogs was explored (three

bogs were excluded due differences in the method of reporting of vegetation types, which did not

enable comparison of vegetation cover data). Principle components 1 and 2 were plotted, and

distances across Euclidean space interpreted from raw data. Bog size and principle components were

plotted to examine if principle components corresponded to bog size. Average proportional cover (±

SE) of each vegetation type was plotted to examine the relative importance of each vegetation type.

A habitat selectivity analysis (Manly et al. 2002) was used to examine M. fuscus preference for

specific habitat types. We assumed that vegetation preferences would not vary between bogs, so

data across all bogs was combined to improve sample sizes for each vegetation category. A Type I

(Manly et al. 2002) selectivity analysis was conducted in the R package AdehabitatHS (Calenge 2013).

The proportion of available habitat was calculated as the proportion of quadrats containing each

vegetation type, and M. fuscus usage was estimated as the count of quadrats that contained M.

fuscus scats corresponding to each vegetation type. Habitat selection indices (Wi) were calculated for

each vegetation type. Overall significance of habitat selection was determined by the log-likelihood

statistic (Manly et al. 2002).

A total of three additional observers assisted the lead author in identifying M. fuscus scats at five

bogs (Cheyenne, Ginini East, Leura, Snowy Flat and Tom Gregory). Only one additional observer was

present at any one time. Observer bias was explored by examining the relative proportion of M.

fuscus scats discovered by an observer, after controlling for bog-level differences in M. fuscus

abundance. If observer bias was present, it would be expected to manifest itself through differences

in M. fuscus scat detection between observers. M. fuscus scat abundance was modelled against

observer, with bogs included as a random factor to account for bog-level differences in M. fuscus

abundance.

The relationship between proportion of quadrats with M. fuscus scats and distance (transformed by

the square-root to meet assumptions of statistical normality) from nearest drainage line was

explored in a GLMM framework. Bog ID was included as a random factor, to account for possible

differences between bogs. In a similar manner, the relationship between M. fuscus abundance and

incidence of feral animal disturbance was explored in a mixed statistical model. All feral animal

disturbance indicators (fox, rabbit and pig) were combined into a single category. While it is plausible

that the relationship between M. fuscus and these species differs due to ecological relationships

(competition, predation), we chose to combine these categories due to limited data for each feral

species. Mixed effects models were implemented in R with the package lme4 (Bates et al. 2011).

Plots were produced in R, with the fixed factors plotted for each level of the random factor (bog ID).

A final model, fitting M. fuscus abundance against the three positively selected vegetation types,

feral animal disturbance and square-root of linear distance to drainage (m). Bog ID was included as a

random factor. The presence of spatial autocorrelation in the model residuals was explored by way

of Moran’s I, in the R package ‘ape’ (version 3.2, Paradis et al. 2003). No significant spatial

autocorrelation was detected (P = 0.84). Model fitting was performed in R (version 2.13; R

Development Core Team, 2011). Model selection was conducted in ‘glmulti’ (Calcagno and de


Mazancourt 2010), using an exhaustive screening of all possible models with 2-way interactions. The

best model was identified using AIC. Figures were produced in Sigmaplot (version 10; Systat Software

Inc., London).

Map 1. Survey design and site establishment for Mastacomys fuscus in the ACT (Snowy Flat).


4 Results

4.1 Bogs

Fourteen alpine bogs in Namadgi National Park were surveyed for M. fuscus scats and vegetation

characteristics in March, 2013. A total of 13 vegetation/habitat types were recorded from quadrats

across all bogs. A Principle Components Analysis revealed that the proportion of vegetation type

could provide separation between bogs in Euclidean space, with PC1 and PC2 explaining 46.39% and

20.05% of the total variation in vegetation between bogs respectively (Figure 1). Proportion of

vegetation in the categories of Exotic plants, Empodisma, HeathL, Sphagnum, Richea, and HeathT

were particularly informative in describing differences between bogs, whereas the proportion for

categories of Poa, rocks, Forbs and woodland were not (Figure 1a). Tom Gregory and Cotter Gap bogs

differed from the other bogs due to their relatively high presence of exotic species (33% and 16.7%,

respectively), while Little Bimberi, Cheyenne Flat and Ginini differed from the other bogs in their

relatively higher proportions of Sphagnum and Richea (Figure 1b). Poa was the most abundant

vegetation type in 11 of the 14 bogs, being present in 75% ± 3.63 SE of all quadrats (Figure 2).

HeathL, HeathM, Empodisma and Sedge were present in more than 20% of quadrats on average,

across all bogs (Figure 2). Conversely, the categories of Forbs, Richea and Juncus, and rocks were

relatively rare in bogs, each being present at less than 5% of all quadrats. The abundance of exotic

plant species was relatively low in bogs; this category was recorded at just three bogs (Tom Gregory,

Cotter Gap and Murray Gap), with bog Tom Gregory having the highest proportion of quadrats

containing exotic plants (33%).

4.2 Distribution of M. fuscus

Across all bogs, M. fuscus scats were found in 26.5% of quadrats examined. At the landscape-scale,

M. fuscus density (proportion of quadrats within a bog that had M. fuscus scats) varied greatly

between bogs (Figure 3a). M. fuscus density was greatest at Cotter Gap bog (69%), and none were

present at Cotter Source bog (0%; Table 1; Figure 3a). The probability of detecting M. fuscus scats

was not related to bog size, however when excluding an outlier (bog Cotter Gap), which supported

an unusually high density of M. fuscus, there was a significant positive relationship between bog size

and M. fuscus density (Figure 3b). PC1 was negatively correlated to bog size, suggesting that

vegetation types vary by bog size (Figure 4). However, M. fuscus abundance was not related to PC1

or PC2.

4.3 Habitat preference of M. fuscus

Pooling data across all bogs, habitat selectivity indices (Wi) suggested that M. fuscus show

preferences for specific vegetation types (log-likelihood statistic: X2 = 31.75, df = 12, P < 0.001)

(Figure 5). Specifically, after bonferroni correction for multiple comparisons, M. fuscus scat presence

was found to be significantly positively associated with medium height heath (HeathM) (Figure 5;

Appendix Figure C). M. fuscus also showed weak but non-significant preferences for Poa and Sedge

(Appendix Figure D-E). Conversely, there was a weak, non-significant negative association with

Empodisma, woodland and low heath (HeathL) (Appendix Figure F-G). Apparent preferences for

rarely occurring vegetation types (such as Forbs, Richea, Juncus) and rock, were given a low

confidence due to the low abundance of these vegetation types and hence small sample sizes. The


probability of M. fuscus scats being observed was negatively related to straight-line distance from

drainage lines, with the likelihood of detecting M. fuscus scats declining by 1.1% per m distance from

a drainage line (Figure 6; Table 2). This relationship was found to be reasonably consistent between

bogs, after accounting for the high bog-level variance (59%) (Table 2; Figure 6).

4.4 Feral animal disturbance

The probability of detecting M. fuscus scats decreased significantly if evidence of feral animal

disturbance was present (Table 3). Feral animal disturbance was observed in 11.2% of quadrats

across all bogs, with Cotter Source having the highest incidence of feral animal disturbance (31.1%;

Table 1). Of the observed feral animal disturbance in bogs, evidence of pigs was most common

(80.2%), with fox and rabbit representing 18.1% and 1.7% respectively. With one exception,

modelling revealed a consistent decrease in M. fuscus density at the bog level, suggesting that feral

animals have a consistent, negative effect on M. fuscus abundance. The exception occurred at Cotter

Gap, where M. fuscus density increased with feral animal disturbance (Figure 7).

A model explaining M. fuscus scat presence with respect to the three most preferred vegetation

types, distance from drainage and presence of feral animal disturbance was found to provide the

best fit, compared to models with a subset of these variables (Table 4). Likelihood of detecting M.

fuscus scats declined with distance from drainage line, and presence of feral animal disturbance.

Conversely, likelihood of detection increased significantly with the presence of Poa, Sedge and

medium height heath (Table 4). All interaction terms were found to be non-significant when the main

effect was present within the model. The random factor (bog ID) explained 34% of variation.

4.5 Observer bias

Observer bias was found to be non-significant. After accounting for bog-level variance, there were no

significant differences in scat detection rates among observers (Table 5). This is evidenced by the

minor differences between observer coefficients, relative to standard errors (Table 5).


Figure 1. Principle component analysis of 13 vegetation types recorded across 11 alpine bogs. a) Ordination plot of principle component 1 and principle component 2, explaining 46.39% and 20.05% of variation in vegetation types, respectively, and b) vector plot, showing the relationships between bogs based upon similarities and differences in vegetation types.


Figure 2. Average percentage of each vegetation type present at each site along transects for 11 alpine bogs. Poa was the most common vegetation type, followed by Empodisma and medium heath.

Table 1. Summary of survey effort, M.fuscus detection rate and feral animal disturbance per bog.

Bog Transects Plots Plots occupied Detection rate %

Feral animal disturbance %

Cotter Gap 10 42 29 69 17

Cheyenne 11 77 33 43 4

Rock Flat 8 70 24 34 13

Snowy Flat 11 133 44 33 6

Ginni West 12 100 30 30 13

Big Creamy 10 84 21 25 10

Ginni East 8 131 31 24 18

Tom Gregory 11 78 18 23 5

Rotten Swamp 6 72 13 18 8

Little Bimberi 11 61 11 18 11

Murray Gap 10 63 10 16 2

Leura 5 28 4 14 7

Top Flat 10 53 7 13 19

Cotter Source 10 45 0 0 31


Figure 3. a) Proportion of spots that contained M. fuscus scats, by bog ID, and b) relationship between index of M. fuscus abundance (from panel a) and bog size (km2). Linear model is fitted to 13 bogs with an outlier (Cg; Cotter Gap) excluded (open circle).


Figure 4. Bog size (km2) plotted against principle component 1 (PC1) from the PCA analysis. An outlier (bottom left; bog Lb; Little Bimberi) was excluded from the linear model. When included, model fit is still significant (F1,9 = 9.37, P = 0.01), although quality of fit is reduced (R2 = 0.51).

Figure 5. Habitat selection index (Wi) for M. fuscus, averaged across 11 alpine bogs. Vegetation types greater than the dashed line show positive selection by M. fuscus, while those below the dashed line are negatively selected by M. fuscus.


Table 2. Summary of general linear mixed effects model on likelihood of detecting M. fuscus scats, and square-root linear distance from drainage, across 14 alpine bogs in Namadgi NP. Bog ID was included as a random factor (random intercept and slope) to account for bog-level variance.

Fixed Factors Coefficient SE Z P

Intercept -0.66 0.245 -2.698 0.007

DistDrainL1/2 -0.088 0.026 -3.38


Table 3. Summary of general linear mixed effects model of likelihood of detecting M. fusucs scats and incidence of feral animal disturbance, across 14 alpine bogs in Namadgi NP. Bog ID was included as a random factor (random intercept and slope) to account for bog-level variance.


Intercept -1.07 0.22 -4.9


Table 4. Summary of general linear mixed effect model on likelihood of detecting M. fuscus scats and key vegetation types (Poa, heathM and sedge), distance from drainage, and incidence of feral animal disturbance, across 11 alpine bogs in Namadgi NP. Bog ID was included as a random factor (random intercept) to account for bog-level differences in mean M. fuscus abundance.


Intercept -1.24 0.3 -4.08


Map 2. ACT Mastacomys fuscus distribution 2013.


5 Discussion

5.1 Distribution, status and abundance

Mastacomys fuscus was detected across 13 of 14 bogs surveyed in the ACT. Detection rate increased

as bog size increased and decreased as distance to drainage line increased. The species showed a

positive habitat preference for heath, sedge and Poa dominated vegetation types. Feral animal

disturbance negatively influenced detection rate in almost all bogs.

The 14 ACT subalpine bogs surveyed in this study ranged in altitude from 1028 m to 1629 m, and

incuded Cotter Gap, Cheyenne, Rocks Flat, Snowy Flat, Ginini West, Big Creamy, Ginini East, Tom

Gregory, Little Bimberi, Rotten Swamp, Murray Flat, Leura and Top Flat. M. fuscus was detected in 13

of these bogs; the only bog surveyed that did not support detectable levels of M. fuscus was the

Cotter Source bog. Feral animal disturbance in this bog was high with over 31% of quadrats showing

some sign of feral animal disturbance. The results of this study increase the number of known sites to

have supported M. fuscus in the ACT from 14 to 23 (Map 2), thereby extending the species’ known

distribution in the region. Previous evidence of the species in the ACT comes from only three surveys

from over 23 years ago and opportunistic observations at some sites of runways and scats. Due to

the nature of the previous surveys (i.e. presence/absence) it is not possible to identify trends in

population size over the past 40 years. The information collected in the current study can be used as

a baseline for longer-term monitoring of M. fuscus distribution, abundance and habitat use and also

for detecting shifts in the composition and distribution of bog vegetation communities, which might

be expected in response to climate change. Shifts in composition and distribution of alpine

vegetation communities due to climate change have been identified as a key threat to the

persistence of M. fuscus across its range (DECC NSW 2007).

5.2 ACT Habitat

M. fuscus was found to occur across a variety of habitat types in the ACT including: rocky outcrops,

tussock grasslands, sedgelands and heathlands, frequently within close proximity to watercourses.

Evidence of woodland habitat use in areas adjacent to subalpine bogs and grasslands was also found,

although results of a habitat selectivity analysis indicated a weak negative association with the

vegetation type, most likely due to a lack of appropriate ground cover. Overall, M. fuscus was

strongly associated with sites dominated by heath (50–150 cm high) and/or areas dominated by Poa

and sedges. Surprisingly M. fuscus showed a weak, negative association with low spreading heath,

which would be expected to provide excellent cover and protection from predators (DECC NSW

2007). However, the accuracy of this result is uncertain because scat detectability may have been

reduced due to poor visibility.

Proximity to drainage lines is apparently important to M. fuscus, with the likelihood of detecting the

species decreasing 1.1% every metre away from a drainage line. These results are consistent with

observations from Kosciuzsko National Park that indicate the species is most frequently encountered

within 15 m of watercourses (K Green pers. comm.). Drainage lines are likely to support a higher

density and complexity of vegetation, as well as a higher abundance of grasses, both of which are key

habitat attributes for the species. M. fuscus feed primarily on monocotyledons and is generally

associated with dense vegetation cover (Carron et al. 1990). Nil spatial autocorrelation in model


residuals indicate that in the present study, linear distance from nearest drainage is an adequate

predictor of M. fuscus distribution in relation to drainages.

5.3 Bog size

The detection rate of M. fuscus increased as bog size increased. In a number of animal species,

habitat patch size has been shown to influence population persistence (Banks et al. 2007). Larger

habitat patches are likely to support larger populations as well as provide a greater ‘target’ for

individuals dispersing into the area (Hanski and Simberloff 1997). Interestingly O’Brien et al (2008)

showed that there is no evidence of patch size influencing population persistence, although habitat

patch sizes for M. fuscus in the Barrington Plateau tended to be larger than those found in the ACT.

There is evidence of populations of M. fuscus in the ACT region declining in isolated swamps, most

likely the result of reduced dispersal success through predation by introduced predators (O’Brien et

al. 2008). M. fuscus are good dispersers and have been shown to recolonise high quality patches

following local extinctions, though long distance movements are uncommon, especially if riparian

cover or creek lines are absent (Bubela et al. 1991; Happold 1995; O’Brien 2008). Drainage systems

between bogs should therefore be managed as a key habitat resource for the species, aiding

dispersal and thereby the persistence of the species in the region.

5.4 Vegetation composition

Vegetation composition varied with bog size. Across all bogs surveyed, Poa grass was the most

abundant species recorded, followed by heath (


association with both Poa grasslands and sedgelands and negative association with sphagnum.

However, with elevated summer temperatures, woody vegetation is likely to increase in alpine areas

(Williams and Costin 1994). M. fuscus were shown in this study to utilise woodlands adjacent to high

quality habitat, though results of the habitat selectivity analysis showed that M. fuscus tend to avoid

woodland habitat and therefore the expansion of woody vegetation may reduce available optimal

habitat and lead to an overall reduction in habitat complexity and structure. Furthermore, while

restricted altitudinal movement is recognised as a significant impact of climate change for M. fuscus

populations in the Barrington Plateau (NSW), the significance of such processes in ACT populations

are currently unknown. Populations of M. fuscus have been identified across a range of altitudinal

gradients (1028 m to 1629 m) and, unlike the Barrington Plateau population, altitudinal movement

for the species in the ACT is less restricted. Indeed, populations of M. fuscus have been identified at

relatively low altitudes in the ACT (e.g. 1028 m).

Fire

An increase in fire frequency within alpine bog communities is predicted to increase the number of

peaty sedgelands, grasslands and heath, while decrease the number of Sphagnum peatlands

(MacDonald 2009). While an increase in sedgeland, grassland and heath vegetation is likely to

increase available habitat for M. fuscus in the long-term, high intensity fire is likely to significantly

impact populations of M. fuscus in periods immediately following fires through direct mortality and

loss of resources (e.g. food and shelter). Even in unburnt areas, predation by foxes on M. fuscus and

other small mammals has been shown to increase following fires, with predation concentrated in

areas still carrying populations of small mammals (Green and Sanecki 2006). The wildfire in 2003

burnt most of the habitat for M. fuscus, though individuals managed to persist in unburnt habitat

patches as evidenced by recent scats and runways in these patches (Carey et al. 2003). While M.

fuscus are good dispersers and have been shown to recolonise high quality patches following

extinction (i.e. following high intensity fires), re-colonisation following such events will depend on the

level of connectivity between populations as well as proximity from source populations. While the

Top Flat population is relatively isolated from other known populations of M. fuscus, the population

is well connected to other populations via suitable dispersal habitat along the Cotter River.

Nevertheless, actions to reduce high intensity fire risk should be incorporated into the management

of such relatively isolated sites.

It should be noted that a number of heath species associated with M. fuscus require fire for

reproduction. A low intensity fire frequency of 10–40 years is likely to be appropriate for maintaining

or improving heath habitat (NPWS 2005) for M. fuscus. Sphagnum peatlands (an Endangered

Ecological Community under the EPBC Act 1999) and surrounding woodland habitat should not be

subject to this fire regime, and instead fire should be excluded from these habitats, particularly the

fire-sensitive Sphagnum moss bog communities. While peatlands in good condition are relatively

resistant to low intensity fire, increased evaporation and reduced rainfall with a changing climate

(Lucas et al. 2007) is predicted to dry out Sphagnum peatlands and significantly increase the risk and

impact of fire within the community (Macdonald 2009).

Habitat degradation through feral animal disturbance

Habitat degradation by feral pigs and rabbits poses a significant threat to populations of M. fuscus in

the ACT. Damage caused by rabbits and pigs was recorded in all 14 bogs surveyed and was

significantly correlated with lower M. fuscus detection rate across all bogs except Cotter Gap. Feral


pig and rabbit activity is likely to result in erosion, a reduction in food resources and a change in

habitat structure and composition (Green and Osborne 2003).

Feral animal activity is likely to contribute significantly to the establishment and spread of weeds and

it is possible that feral pigs could play a key role in the spread of Phytophthora. Phytophthora poses a

significant threat to the species habitat – leading to a loss of heath and other vegetation essential to

the species survival. Bogs with M. fuscus and other threatened species should be priority target areas

for landscape scale feral animal control programs in Namadgi National Park.

Weeds

The invasion of exotic plant species into key M. fuscus habitat poses a significant threat to the

species survival. The invasion of exotic grasses is likely to result in the loss of important ground layer

vegetation such as Snow grass with the potential to reduce the availability of that food source for M.

fuscus. Yorkshire fog Holcus lanatus was recorded at relatively high density at the Cotter Gap bog,

though there is currently no evidence to indicate that this introduced species is having an adverse

affect on the population as M. fuscus detection rate in this bog was high. H. lanatus has, however,

been identified as a potential threat to populations of M. fuscus in the Barrington Tops where there

is concern that this grass may displace significant native ground layer species such as Snow Grass,

which M. fuscus is known to rely heavily on as a food resource (DECC NSW 2007). Careful monitoring

of the Cotter Gap bog, and other bogs that can be accessed by the public on foot or by vehicle, is

recommended to ensure that infestations of H. lanatus and other weeds do not increase and that the

presence of the species is not having an adverse affect on the population of M. fuscus. It should be

noted that H. lanatus is unlikely to be a food resource for M. fuscus.

Competition

A number of both exotic and native species compete with M. fuscus for food, shelter and space.

Evidence of rabbits was recorded in nine of the 14 sites surveyed, with disturbance highest at Cotter

Gap, Cotter Source and Tom Gregory. Rabbits compete directly with M. fuscus for food; and grazing

by rabbits can result in the removal or modification of tussock structure that provides shelter and

nesting habitat for M. fuscus. Control of rabbits and foxes is likely to be most effective when

undertaken concurrently at a landscape scale, and control of these species is particularly important

at the following three bogs: Cotter Source, Cotter Gap and Tom Gregory.

Competition for resources from native species, such as the Bush Rat Rattus fuscipes, may pose an

additional competitive threat to M. fuscus. For example, a significant increase in the number of R.

fuscipes trapped in Nolans Swamp (NSW) coincided with a dramatic decrease in the number of M.

fuscus captured, indicating a potential dominance of R. fuscipes over M. fuscus (Keating 2003). R.

fuscipes were detected in four of the 14 bogs surveyed (incl. Cheyenne, Murray Gap and Ginini East

and West) in the ACT. However, R. fuscipes is naturally sympatric with M. fuscus and therefore it is

unknown if and under what circumstances R. fuscipes poses a potential competitive threat to M.

fuscus in ACT populations.

Predation

M. fuscus fall within the Critical Weight Range (CWR; Burbidge and McKenzie 1989) and the

Vulnerable Terrestrial Vertebrate category (Smith and Quin 1996) indicating a high risk of population

decline and extinction through predation from foxes and cats. There is also evidence of selective


predation by foxes on M. fuscus over other sympatric CWR rodent species. For example, Green

(2002) showed that M. fuscus in the alpine regions of the Snowy Mountains were five times more

frequently found in fox scats than were R. fuscipes despite occurring at relatively similar densities.

The disparity in predation rate has been attributed to behavioural difference. For example, M. fuscus

are less aggressive than R. fuscipes, they are slower, they use well-established runways and they nest

communally above the ground during winter.

Evidence of foxes was found in two of the 14 bogs surveyed, including Ginini West and Cotter Source.

5.6 Conclusion

M. fuscus is present at a range of sites in the higher elevation areas of the ACT where there is

suitable habitat. Whilst the wildfire of 2003 burnt most of the habitat for the species, individuals

survived in unburnt patches, and 13 years post-fire the species appears to still maintain its pre-fire

distribution in the ACT. To what extent the 2003 fires reduced populations of M. fuscus in the ACT

and resulted in the species going through a genetic ‘bottleneck’ is not known. Although it is not

possible to determine a trend in abundance from a single ‘snapshot’ study, the high occupancy rate

across the majority of bogs surveyed in the ACT subalpine region suggests that the populations of the

species are currently relatively stable, albeit within a small and patchy distribution. On-going

monitoring is required to determine whether ACT populations of M. fuscus are experiencing longer-

term declines that appear to have occurred in other areas. This study provides baseline information

on vegetation parameters and an index of density of M. fuscus and of introduced predators and

competitors in bogs against which future changes can be assessed and key-threatening processes can

be monitored. Population monitoring of M. fuscus and vegetation assessments undertaken every 5-

10 years would enable long-term trends to be identified. The methods used in this study are an

efficient and repeatable method for identifying trends in M. fuscus abundance and distribution.

Acknowledgements

We thank Melissa Snape, Luke Johnston and Sophia Callander for assistance in the field. We also

thank Melita Milner, Felicity Grant, David Whitefield, Ken Green and Will Osborne for assistance.

References

Banks, S., Piggott, M.P., Stow, A.J. and Taylor, A.C. 2007. Sex and sociality in a disconnected world: a

review of the impacts of habitat fragmentation on animal social interactions. Canadian Journal of

Zoology 85, 1065-1079.

Bates, D., Maechler, M. & Bolker, B. 2011. Lme4: Linear Mixed-Effects Models using S4 Classes.

Brereton, R., Bennett, S. and Mansergh, I. 1995. Enhanced greenhouse climate change and its

potential effect on selected fauna of south-eastern Australia: a trend analysis. Biological

Conservation 72, 339-354.

Bubela, T.M., Happold, D.C.D. & Broome, L.S. 1991. Home range and activity of the Broad-toothed

rat, Mastacomys fuscus, in Subalpine Heathland. Wildlife Research 18, 39-48.


Bubela, T.M. & Happold, D.C.D. 1993. The social organization and mating system of an Australian

sub-alpine rodent, the Broad-toothed rat, Mastacomys fuscus (Thomas). Wildlife Research 20,

405-417.

Calaby, J.H. & Wimbush, D.J. 1964. Observation on the broad-toothed rat, Mastacomys fuscus

Thomas. Wildlife Research 9, 123-133.

Calcagno, V. & De Mazancourt, C. 2010. glmulti: An R package for easy automated model selection

with (generalized) linear models. J Stat Soft 34, 1–29.

Calenge, C. 2013. Analysis of habitat selection by animals; Package: adehabitatHS.

Carey, A., Evans, M., Hann, P., Lintermans, M., MacDonald, T., Ormay, P., Sharp, S., Shorthouse, D.

and Webb, N. 2003. Technical Report 17 Wildfires in the ACT 2003: Report on initial impacts on

natural ecosystems. Environment ACT, Canberra.

Carron, P.L., Happold, D.C.D. & Bubela, T.M. 1990. Diet of two sympatric Australian Subalpine

Rodents, Mastacomys fuscus and Rattus fuscipes. Australian wildlife Research 17, 479-489.

Department of Environment and Climate Change (NSW). 2007. Draft Recovery Plan for the Barrington

Tops Broad-toothed Rat Endangered Population, Department of Environment and Climate Change

(NSW), Sydney.

Eberhard, I. & Schulz, L. 1973. A survey of the vertebrate fauna of the Cotter River Catchement,

Australian capital Territory. Department of the Capital Territory, conservation Memorandum No

10.

Green, K. 2000. A survey of the Broad-toothed Rat at Barrington Tops, Report to the NSW, NPWS,

unpub.

Green, K. 2002. Selective predation on the broad-tooted rat, Mastacomys fuscus ( Rodentia:

Muridae), by the introduced red fox, Vulpes vulpes (Carnivora: Canidae), in the Snowy Mountains,

Australia. Austral Ecology 27, 353-359.

Green, K. & Osborne, W.S. 2003. The distribution and status of the Broad-toothed rat Mastacomys

fuscus (Rodentia: Muridae) in New South Wales and the Australian Capital Territory. Australian

Zoologist 32, 229-237.

Green, K. and Sanecki, G. 2006. Immediate and short term responses of birds and mammal

assemblages to a subalpine wildfire in the Snowy Mountains, Australia. Austral Ecology 31, 673-

681.

Hanski, I., and Simberloff, D. 1997. The metapopulation approach, its history, conceptual domain,

and application to conservation. In ‘Metapopulation biology’. (Ed in I. Hanski and M. Gilpin). Pp. 5-

26. (Academic Press, San Diego, California, USA).

Happold, D.C.D. 1989b. The value of faecal pellets for ascertaining the presence of Mastacomys

fuscus (Rodentia, Muridae) in field surveys. Victorian Naturalist 106, 41-43.

Happold , D.C.D. 2002. ‘Broad-toothed Rat’, in The Mammals of Australia, ed R. Strahan, Reed Books,

Sydney.

Happold, D.C.D. 2011. Reproduction and ontogeny of Mastacomys fuscus (Rodentia: Muridae) in the

Australian Alps and comparison with other small mammals living in alpine communities.

Mammalian Biology 76, 540-548.


Keating, J. 2003. Surveys for the Broad-toothed Rat in Barrington Tops National Park, Report to NSW

NPSW, Nelson Bay.

Lucas, C., Hennessy, K., Mills, G. & Bathols, J. 2007. Bushfire weather in southeast Australia: recent

trends and projects climate change impacts. Consultancy report prepared for the Climate Institute

of Australia, Bushfire Cooperative Research Centre, Melbourne.

Lintermans, M. 1986. The Broad-toothed Rat Mastacomys fuscus in the ACT. ACT Parks and

Conservation Services.

Macdonald, T. 2009. Technical Report 20. Sphagnum Bog Mapping and Recovery Plan. ACT Climate

Change Strategy Action Plan 2007-2011, Action 35 Project Report. Territory and Municipal

Services, Canberra.

Manly, B.F.J., McDonald, L.L., Thomas, D.L., McDonald, T.L. & Erickson, P. 2002. Resource Selection by

Animals: Statistical Design and Analysis for Field Studies, 2nd edn. Kluwer Academic Publishers,

London.

McDougall, K.L. and Broome, L.S. 2007. Challenges facing protected area planning in the Australian

Alps in a changing climate. In ‘Protected Areas: Buffering nature against climate change’.

Proceedings of a WWF-Australia and IUCN World Commision on Protected Areas symposium,

Canberra 18-19 June 2007. (Ed M Taylor and P Figgis). Pp. 73-84. (WWF Australia, Sydney).

Milner R, Starrs D, Hayes G and Evans M. 2014. Distribution and habitat preference of the broad-

toothed rat (Mastacomys fuscus) in the Australian Capital Territory, Australia. Australian

Mammalogy. http://dx.doi.org/10.1071/AM14031

O’Brien, C.M., Crowther, M.S., Dickman, C.R. & Keating, J. 2008. Metapopulation dynamics and

threated species management: Why does the broad-toothed rat (Mastacomys fuscus) persist?

Biological conservation 141, 1962-1971.

Paradis, E., Claude, J. and Strimmer, K. 2003. Ape: Analyses of phylogenetics and evolution in R

language. Bioinformatics 20, 289-290.

R Development Core Team (2011) R: a language and environment for statistical computing. Vienna,

Austria

Ride, W.D.L. 1956. A new fossil Mastacomys (Muridae) and a revision of the genus. Proceedings of

the Zoological Society of London 127: 431-439.

Williams, R.J. & Costin, A.B. 1994. Alpine and subalpine vegetation. In Australian Vegetation (ed R.H.

Groves) pp. 467-500. Cambridge University Press, Cambridge.

Personal Communications

Murray Evans, Conservation Research section, ACT Government.

David Whitfield, Parks and Conservation Service, ACT Government.

Ken Green, Office of the Environment and Heritage, NSW Government.


Appendix

Figure A. Mastacomys fuscus runway.


Figure B. Mastacomys fuscus scats.

Figure C. Vegetation community dominated by medium heath.


Figure D. Vegetation community dominated by Poa.

Figure E. Vegetation community dominated by sedge.


Figure F. Woodland vegetation community.

Figure G. Vegetation community dominated by low heath.

distribution and ecology of the broad-toothed rat in …€¦ · determine the distribution,...

Documents